https://github.com/smallhelm/sdr

javascirpt functions to work with Sparse Distributed Representations
https://github.com/smallhelm/sdr

Last synced: 7 months ago
JSON representation

javascirpt functions to work with Sparse Distributed Representations

• Host: GitHub
• URL: https://github.com/smallhelm/sdr
• Owner: smallhelm
• License: mit
• Created: 2016-06-30T02:19:40.000Z (almost 10 years ago)
• Default Branch: master
• Last Pushed: 2016-06-30T03:48:40.000Z (almost 10 years ago)
• Last Synced: 2025-01-11T07:38:52.937Z (over 1 year ago)
• Language: JavaScript
• Homepage:
• Size: 4.88 KB
• Stars: 4
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# sdr

[![build status](https://secure.travis-ci.org/smallhelm/sdr.svg)](https://travis-ci.org/smallhelm/sdr)

Functions for working with [Sparse Distributed Representation](https://github.com/numenta/nupic/wiki/Sparse-Distributed-Representations)

> one of biggest problems in AI

> --no, the only problem in AI--

> is the problem of representation

So said a [retiring AI researcher](https://youtu.be/A8sHMcCk0lU?t=9m51s)

If you don't know what SDRs are, you should go learn about them. [numenta.org](http://numenta.org) Or watch [these videos](https://www.youtube.com/playlist?list=PL3yXMgtrZmDqhsFQzwUC9V8MeeVOQ7eZ9) for an easy to follow introduction.

## Usage

Your SDR arrays are always assumed to be in order from least to greatest. The functions use this assumption to optimize performance.

```js

var sdr = require('sdr');

sdr.union(

  [0, 10, 15, 30],

  [0,  5, 10, 20]

);

//-> [0, 5, 10, 15, 20, 30]

sdr.intersect(

  [1, 2],

  [2, 3]

);

//-> [2]

sdr.sparsity([0, 10, 20], 100);

//-> 0.03

sdr.capacity(2048, 40);

//-> 2.3717785116453587e+84

```

## License

MIT